Research Article

Comparison Between the Effect of Strongly and Weakly Cationic ExchangeResins on Matrix Physical Properties and the Controlled Releaseof Diphenhydramine Hydrochloride from Matrices

Prasert Akkaramongkolporn,1,3 Kaewnapa Wongsermsin,2 Praneet Opanasopit,1 and Tanasait Ngawhirunpat1

Received 7 March 2010; accepted 11 June 2010; published online 9 July 2010

Abstract. This study focused on investigating and comparing between the effect of the strongly cationicexchange resin, Dowex 88 (Dow88), and the weakly cationic exchange resin, Amberlite IRP64 (Am64),on the physical properties of matrices and their drug release profiles. The matrices were prepared bydirect compression of Methocel K4M (HPMC) or Ethocel 7FP (EC) polymeric matrix formers andcontained diphenhydramine hydrochloride as a model drug. The addition of Dow88 to the matricesdecreased matrix hardness and increased thickness, diameter, and friability. In contrast, the addition ofAm64 increased matrix hardness and maintained the original thickness, diameter, and friability. Indeionized water, both resins lowered drug release from HPMC-based matrices by virtue of the gelationproperty of matrix former and the drug exchange property of embedded resin, in other words in situresinate formation. Dow88 strongly dissociated and lowered the drug release to a greater extent thanAm64, which was weakly dissociated. However, Am64 could retard drug release under simulatedgastrointestinal conditions. EC-based matrices containing either resin displayed a propensity fordisintegration caused by swelling and wicking (water adsorption) actions by the resin. The results ofthis study provided useful information on the utilization of ion exchange resins as release modifiers inmatrix systems.

KEY WORDS: controlled release; matrices; physical properties; strongly cationic exchange resin; weaklycationic exchange resin.

INTRODUCTION

Ion exchange resins are insoluble, crosslinked copoly-mers that can reversibly bind dissociated drugs via ionexchange. Resins have diversified applications in chemicaland pharmaceutical industries. In pharmaceutics, they areoften used as drug carriers and to control the rate of drugrelease. Controlled release systems using resin technology canbe administered in a variety of forms including tablet,capsule, powder, gum, and liquid suspension. The bitter andunpleasant taste of drugs is reduced when bound to resins;thus, patients are more compliant in consuming these “resin-ates.” Furthermore, resins can be exploited as a disintegrantin tablet formulations (1–3).

Matrix systems, especially those prepared by directioncompression, are a popular method for controlling the releaseof drugs (4). Materials used for forming matrices includehydrophobic or hydrophilic polymers. Drug release frommatrices using hydrophobic polymers such as ethyl celluloseproceeds via diffusion through an almost intact matrix (5,6).

Matrices made of hydrophilic polymers such as hydroxy-propyl methylcellulose swell and form gels that may simulta-neously erode if the molecular weight of the polymer is low.Nonetheless, the release of water-soluble drugs is governedprimarily by diffusion rather than erosion of gelled matrices(7,8). Common methods of modifying drug release frommatrices include selecting a suitable polymer, adjusting thepolymer/drug ratio, admixing several polymers, and addingan excipient such as a pore former (4–9).

Recent studies have shown that resins can modify therelease of drugs from matrices without the need for priorformation of a resinate. Thus, drug release may be governedby “in situ resinate formation” (10,11). Adding an ionexchange resin to modify the drug release may alter thephysical properties of matrices such as hardness and friability(11). The literature suggests that strongly cationic exchangeresins such as Amberlite IRP69 and Dowex 50W can retarddrug release from matrices and adversely affect matrix hard-ness and friability. However, the mechanism of drug releasemodification and the effect of weakly cationic exchange resinson the physical properties of matrices have not been estab-lished. This information is necessary for adopting ion exchangeresins as a release modifier for matrix systems.

Thus, the objective of this study was to investigate theeffect of the weakly cationic exchange resin, Amberlite IRP64(Am64), on the physical properties of matrices and their drug

1Department of Pharmaceutical Technology, Faculty of Pharmacy,Silpakorn University, Nakhon Pathom 73000, Thailand.

2 Faculty of Pharmacy, Siam Universiy, Bangkok Thailand.3 To whom correspondence should be addressed. (e-mail: prasert@su.ac.th)

AAPS PharmSciTech, Vol. 11, No. 3, September 2010 (# 2010)DOI: 10.1208/s12249-010-9472-2

1530-9932/10/0300-1104/0 # 2010 American Association of Pharmaceutical Scientists 1104

release profiles. In addition, the effect of the strongly cationicexchange resin, Dowex 88 (Dow88), on the matrix propertiesand drug release was investigated and compared to those ofAm64. The matrices were prepared from either MethocelK4M hydroxypropyl methylcellulose (HPMC) or Ethocel 7FPethyl cellulose (EC), which represented hydrophilic andhydrophobic polymeric matrix formers, respectively. In thisstudy, the antihistamine diphenhydramine hydrochloride waschosen as a model drug. Diphenhydramine hydrochloride hasa short half-life of approximately 5–6 h and is administeredorally several times daily (12,13). The drug could be preparedin a controlled release matrix to provide improved conven-ience and patient compliance. Moreover, this drug was usedin our previous work; the use of diphenhydramine hydro-chloride in the current study allowed for a direct comparisonto prior results (11).

MATERIALS AND METHODS

Materials

Methocel K4M hydroxypropyl methylcellulose (viscosityof 3,000–5,600 cps as a 2% solution in water and methoxyland hydroxypropyl contents of 19–24% and 7–12%, respec-tively) and Ethocel 7FP ethyl cellulose (viscosity of 6–8 cps asa 5% solution in 80% toluene and 20% alcohol and a ethoxylcontent of 48.0–49.5%) were kindly donated from ColorconLtd., UK (Dow Chemical Co., USA). Magnesium stearate(BP grade) was a gift from Glaxo Wellcome Vidhyasom,Thailand. Dowex 88 (Dow Chemical Co., USA), AmberliteIRP64, Amberlite IRP69, and diphenhydramine hydrochlor-ide (Sigma Chemical Co., USA) were purchased from varioussuppliers. Other chemicals employed in the investigation wereanalytical grade or higher. Deionized water was used entirelyin this work.

Methods

Preparation of Matrices by Direct Compression

Fixed weight matrices were composed of diphenhydr-amine hydrochloride, a resin (Dow88 or Am64), magnesiumstearate, and a matrix former (HPMC or EC). The drugcontent and lubricant were fixed at 30% and 1% w/w,respectively. The amount of resin was varied from 0% to40% w/w, and the remainder of each matrix was composed of(percent w/w) matrix former. All required components weregently blended in a mortar for 10 min. Then, a total of 100 mgof each mixture was accurately weighed and carefully placedinto a hydraulic hand press machine (Specac P/N 15011/25011, UK). The matrices were compressed using stainlesssteel flat-circular punches (6.35 mm in diameter) with aconstant force and dwelling time of 5 t for 5 s. The producedmatrices were stored tightly in containers until use.

Thickness, Diameter, Hardness, and Friability of Matrices

The thickness and diameter of ten matrices weremeasured with a micrometer (Sylvac S229, Switzerland), andhardness was measured with a texture analyzer (Stable MicroSystems TA.XT plus, UK). To this end, matrices were pressed

by a stainless steel flat-face (6 mm in diameter) cylindricalprobe moving at a constant speed (1 mm/s). The hardnessvalue was read directly from the instrument and was given asthe force that caused a diametrical break in the matrices.Friability was tested on a Roche friabilator. Twenty matriceswere weighed (W1) and placed into the friabilator operatingat 100 rev for 4 min. The matrices were then weighed (W2)again, and the friability (percent) was calculated by theequation: 100� W1 �W2ð Þ W1= (14).

Drug Release from Matrices

(a) In deionized water

Drug release was investigated in triplicate using a USPdissolution testing apparatus I (Dissolutest Prolabo, France)and 900 ml of deionized water at 37±1°C (15). The releasetest was conducted under a constant rotation (50 rpm). Atpredetermined times, 5 ml of the medium was withdrawnthrough a filter, and an identical volume of fresh medium wasadded to the vessels to maintain a constant volume. Thewithdrawn medium was assayed by an ultraviolet spectropho-tometer (UV, Perkin Elmer Lambda 2, Germany) at 218 nm.

(b) Under simulated gastrointestinal conditions

A simulated gastrointestinal release study was conductedfollowing the USP dissolution test for delayed release(method A) using an identical apparatus (15). Drug releasewas determined in 750 ml of a 0.1-N hydrochloric acidsolution (HCl) for 2 h. Thereafter, 250 ml of a 0.2-M tribasicsodium phosphate solution (Na3PO4) was added into thevessels, and drug release into the buffered solution (pH 6.8)was determined. The rotation rate and temperature weremaintained at 50 rpm and 37±1°C, respectively. At prede-termined times, 5 ml of medium was withdrawn through afilter, and an identical volume of fresh medium was added tothe vessels to maintain a constant volume. The withdrawnmedium was assayed by UV at 218 nm, and appropriateblanks were used as correction factors. Drug release wasevaluated in triplicate.

Disintegration Test

Disintegration tests were conducted with a USP disinte-gration testing apparatus (Sotax DT3, Switzerland) (15). Atthe beginning of each test, six matrices were placed into abasket-rack assembly with disks and deionized water at 37±1°C. The disintegration time, defined as the moment thematrix disintegrated and passed through the assembly screen,was recorded for each sample.

Scanning Electron Microscopy (SEM)

The surface morphology of produced matrices wasdetermined with a scanning electron microscope (SEM; JeolJSM 5400, Japan) under a fixed magnification. Prior toviewing, matrices were firmly held on the top of stubs andsputter coated with gold in a vacuum evaporator (SPI ModuleCarbon Coater, USA).

1105Comparison Between Strongly and Weakly Cationic Exchange Resins

Photograph of Matrices During Drug Release

Photographs of matrices were taken under conditionsidentical to the release tests. At predetermined times,matrices were viewed with a digital camera in associationwith image analysis software (Dino-Lite Digital AM-313TPlus, Taiwan) under a fixed magnification.

Swelling of Resin

(a) Volume swelling

Equal weights of resin were placed into a 10-ml cylinderand were tapped until constant volume. The volume of resin

Fig. 1. Chemical structure of a Dow88 and b Am64, respectively

Fig. 2. Hardness (bar graph) and friability (line graph) of a HPMC- and b EC-basedmatrices containing various amounts of resin, respectively

1106 Akkaramongkolporn, Wongsermsin, Opa nasopit and Ngawhirunpat

(v1) was recorded, and an excess amount of water wascarefully added into the cylinder. After 2 h, the volume ofswollen resin (v2) was recorded, and volume swelling(percent) was calculated according to the following equation:100� v2 � v1ð Þ v1= (16).

(b) Weight swelling

Equal weights of resin (w1) were placed in a centrifugetube, and an excess amount of water was carefully added.After 2 h, the slurry was centrifuged at 2,000 rpm until thesupernatant was clear. The supernatant was carefullyremoved with a micropipette, and the swollen resin wasweighed (w2). Weight swelling (percent) was calculatedaccording to the following equation: 100� w2 � w1ð Þ w1=(17).

RESULTS AND DICUSSION

Dow88 and Am64

Dow88 is a strongly cationic exchange resin containingcrosslinked styrene–divinylbenzene copolymer with sodiumsulfonate groups as ion exchange sites (Fig. 1a) (18). Thisresin was milled in a mortar, and the fraction that passed

through a 100-mesh sieve was used. Am64 is a weaklycationic exchange resin containing crosslinked methacrylicacid–divinylbenzene copolymer with carboxylic acid groupsas ion exchange sites (Fig. 1b) (19). Am64 was obtained as amixture of particles ranging in size from 100 to 400 mesh andtherefore was used without modification. The results of anSEM study revealed that both resins were irregular in shape(figures not shown). The resins were insoluble in water butcould hydrate and swell due to the dissociation of ionexchange sites.

Physical Properties of Matrices

All matrices were prepared by direct compression.Matrices with or without resin had a fairly uniform weight(%CV≤1) because mixture ingredients were weighed anddirectly fed into the compress (14). The hardness of HPMC-or EC-based matrices containing Dow88 continuouslydecreased with increasing amounts of embedded resin(Fig. 2), which was in agreement with SEM results. Thedensity of matrix surface deteriorated (Fig. 3a versus b forHPMC-based matrices and Fig. 3d versus e for EC-basedmatrices, respectively) with increasing amounts of embeddedresin. The deterioration in the hardness of HPMC-based

Fig. 3. Surface morphology of HPMC-based matrices containing a no resin, b 40% Dow88, and c 40%Am64 and EC-based matrices containing d no resin, e 40% Dow88, and f 40% Am64, respectively (×50magnification)

1107Comparison Between Strongly and Weakly Cationic Exchange Resins

matrices was higher than EC-based matrices because HPMChad lower compressibility than EC. This result was similar tothose observed from the matrices containing AmberliteIRP69 (Am69) in the previous work (11). Am69 and Dow88are crosslinked styrene–divinylbenzene copolymer resins withsodium sulfonate as ion exchange sites. Am69 is a gel-type(lowly crosslinked copolymer) resin, whereas Dow88 is amacroporous-type (highly crosslinked copolymer) resin(18–20). Thus, hardness results confirmed that crosslinkedstyrene–divinylbenzene copolymer resins had poor compres-sibility regardless of resin type or degree of crosslinking.

Interestingly, the hardness of HPMC- or EC-basedmatrices containing Am64 increased as the amount ofembedded resin increased (Fig. 2). Coincidently, the matricesmaintained smooth and compact surfaces (Fig. 3a versus c forHPMC-based matrices and Fig. 3d versus f for EC-basedmatrices, respectively). This result differed from thoseobserved in matrices containing either Dow88 or Am69(11), which implied that Am64 had higher compressibilitythan Dow88 and Am69. As already mentioned, Am64 is a

crosslinked methacrylic acid–divnylbenzene copolymer resin,whereas Dow88 and Am69 are crosslinked styrene–divinyl-benzene copolymer resins. It was possible that the copolymerchemistry governed the resin compressibility; the methacrylicacid–divinylbenzene copolymer resin might be inherentlymore compressible than styrene–divinylbenzene copolymerresins. Based on this virtue, Am64 was likely to be a superiorrelease modifier for matrices because it did not underminematrix hardness, which was observed with the addition ofDow88 and Am69.

Figure 2 also shows the friability of obtained matrices.The friability of HPMC-based matrices containing Dow88increased markedly and could be divided into two phases(Fig. 2a). In the first phase, friability was gradually increasedfrom 0.3% to 6.0% as the amount of Dow88 increased to20%, corresponding to a decrease in hardness from 47.4 to11.6 N. In the second phase, the friability of matricescontaining 20% to 40% Dow88 increased dramatically from6.0% to 100%, while the hardness slightly declined from 11.6to 4.9 N. In contrast, the friability of EC-based matrices

Fig. 4. Thickness a and diameter b of matrices containing various amounts of resin

1108 Akkaramongkolporn, Wongsermsin, Opa nasopit and Ngawhirunpat

containing Dow88 up to 40% increased slightly (<1.5%),although their hardness continuously decreased from 114.0 to22.6 N (Fig. 2b). Indeed, this result concurred with those ofprevious studies and confirmed that a critical matrix strengthwas present in resin-embedded matrices (11). Below thecritical matrix strength, friability would greatly increase with adecrease in hardness, otherwise remaining nearly unchanged.From Fig. 2, the critical matrix strength of HPMC-basedmatrices containing Dow88 was estimated to be approxi-mately 12 N, which was similar to those of matricescontaining Am69 (16–17 N). Therefore, HPMC-based matri-ces containing 20% to 40% Dow88 had a matrix hardnessbelow 12 N, and the friability greatly increased with decreas-ing the matrix hardness. The friability of EC-based matricescontaining Dow88 was nearly unchanged because their hard-ness had not declined below the critical matrix strength. Inboth HPMC- and EC-based matrices, the friability wasmaintained constant (≤0.3%) when Am64 was added to thematrices (Fig. 2). This result was due to the high compressi-bility of Am64, which increased the hardness of the obtainedmatrices.

The thickness and diameter of the matrices are depictedin Fig. 4. The matrices had similar dimensions; however,HPMC-based matrices containing Dow88 had an increasedthickness and diameter, particularly when the amount ofembedded resin exceeded 10%. This result might be due tothe severe decrease in hardness and inter-particulate bindingof the matrices, which was a result of the poor compressibilityof the original matrix and embedded resin. As mentionedabove, the compressibility of HPMC and Dow88 was lowerthan EC and Am64, respectively. Therefore, Dow88 greatlyweakened inter-particulate binding of the HPMC matrix,leading to a less compact and more relaxed matrix. Due tothe superior compressibility of EC, matrix hardness and inter-particulate binding of EC-based matrices containing Dow88were not lowered as severely as those of HPMC-basedmatrices containing Dow88. Instead, the matrices were strongenough to withhold matrix relaxation after punch withdrawal.As with Am64, it strengthened matrix hardness and inter-particulate binding. As a result, the density of HPMC- or EC-based matrices containing Am64 was maintained, producingan unchanged thickness and diameter.

Drug Release from Matrices in Deionized Water

Drug release from HPMC-based matrices containingDow88 is shown in Fig. 5a. The results revealed that drugrelease was lowered by the incorporation of Dow88 into thematrix, which was similar to the results of previous studies(10,11). The decrease in drug release was due to the gelationproperty of HPMC and the drug exchange property ofembedded resin. Thus, the observed decrease was caused by“in situ resinate formation”; upon contact with water, thematrices formed a gel (Fig. 6) of which the drug was able todiffuse out. However, a portion of the dissolved drug becamebound to the embedded resin via ion exchange and formed aresinate or drug–resin complex, resulting in a decrease indrug release. This decrease in drug release was moredramatic with increasing amounts of embedded resin due toan increase in resinate formation.

The incorporation of Am64 also decreased drug releasefrom HPMC-based matrices (Fig. 5b) due to in situ resinateformation. Matrices containing Am64 formed a gel duringdrug release (Fig. 6); however, the decline in drug releasefrom Am64 was clearly less than Dow88-embedded matricesdue to the weaker drug exchange property of Am64compared to Dow88. The ion exchange sites of Am64 andDow88 are carboxylic and sulfonic groups, respectively, andthe pH of drug solutions was between 5 and 6. Thus, thecarboxylic group of Am64 (pKa=4–6) weakly dissociated andprovided lower drug exchange and in situ resinate formationthan the sulfonic group of Dow88 (pKa=1–2), which couldstrongly dissociate and interact with the dissolved drug(20,21).

It was observed that gelled matrices gradually erodedduring drug release, and the erosion rate increased withincreasing amounts of embedded resin due to a decrease inthe amount of matrix former. For example, gelled matricescontaining 40% Dow88 eroded so much that the small,collapsed residue could not be prepared for photographicviewing at the end of the release test (Fig. 6a). Nevertheless,matrices containing a higher amount of embedded resinprovided a lower drug release in spite of higher matrixerosion. This phenomenon supported that drug release fromthe matrices was primarily governed by drug diffusion and

Fig. 5. DPH release from HPMC-based matrices containing variousamounts of a Dow88 and b Am64, respectively

1109Comparison Between Strongly and Weakly Cationic Exchange Resins

exchange rather than matrix erosion (10,11). Moreover, itcould be further evaluated by the kinetic analysis of releaseprofiles. In general, the release data from gelled matrixsystems can be analyzed according to the power law equation:Ft=kt

n, where Ft is the fractional release of the drug (in therange of 0–70%) at time t, k is the kinetic constant, and n isthe release exponent indicative of the release mechanism,respectively. According to the resultant kinetic parameters

(Table I), all release profiles were satisfactorily fitted to thepower law equation using appropriate n values (R2≥0.98).The value of n from the free-resin matrices was 0.569,indicating an anomalous transport corresponding to drugdiffusion in the hydrated matrix and the polymer erosion(22,23). However, the matrices containing either resinexhibited a trend of decreasing n values, i.e., from 0.569 to0.227 for Dow88 and from 0.569 to 0.410 for Am64,

Fig. 6. Photographs during drug release of HPMC-based matrices containing a 40%Dow88 and b 40% Am64, EC-based matrices containing c 40% Am64 tested in deionizedwater, and HPMC-based matrices containing d 40% Dow88 and e 40% Am64 tested undersimulated gastrointestinal conditions, respectively

1110 Akkaramongkolporn, Wongsermsin, Opa nasopit and Ngawhirunpat

respectively, elucidating that the release mechanism was notpurely dependent on such drug diffusion, but coupled withdrug exchange property from the embedded resin. Thegreater shift of n values was observed from Dow88 due tostronger drug exchange property of the resin. In addition, theeffect of increased amounts of resin and delayed drug releasecould be expressed as slowed kinetic constant (k) for matricescontaining Dow88.

Drug release from EC-based matrices containing Dow88is shown in Fig. 7. During the release test, it was observedthat Dow88 caused non-swelling EC-based matrices todisintegrate with a disintegration rate and extent paralleling

the amount of embedded resin added to the matrix. Thisobservation was further quantitatively confirmed by a dis-integration test. Clearly, disintegration times were shorterwith increasing amounts of embedded resin, which were329.6±4.4, 72.3±1.3, 8.3±0.5, 1.3±0.4, and 0.4±0.1 min forEC-based matrices containing 5%, 10%, 20%, 30%, and 40%Dow88, respectively. A greater disintegration rate allowedthe drug to be released faster (Fig. 7) due to an increase insurface area exposed to the medium (24). However, theamount of drug released into solution tended to be lowerbecause the released drug was bound to resin dispersed in themedium. The decrease in the amount of drug released intosolution was more pronounced as the amount of dispersedresin, and hence, resinate formation in the medium increased.This result was in agreement with previous studies whereAm69 was embedded into EC-based matrices (11).

The mechanism of matrix disintegration due to embed-ded resin was further evaluated. It was believed that resinscaused matrix disintegration via resin swelling; therefore, theswelling of resins was determined (10,11). The swelling ofAm69 was also determined for comparison. Indeed, resultsindicated that the resins swelled volumetrically (volumeswelling) and gravimetrically (weight swelling). The volumeswelling was 19.6±3.1%, 44.0±2.9%, and 68.5±5.2%, whilethe weight swelling was 158.9±14.2%, 231.9±2.2%, and167.2±5.4% for Dow88, Am64, and Am69, respectively. Itcould be found that the volume and weight swelling were notequal and not totally parallel. For example, the volumeswelling of Am64 was lower than that of Am69, but theformer resin had higher weight swelling than the latter resin.This result indicated that volume and weight swelling mighthave different origins. Volume swelling represented anincrease in size (swelling action), whereas weight swellingindicated that water was adsorbed by the resin without achange in volume or size (wicking action) (25). Previousstudies reported that the disintegrating effect of Am69 wasprimarily due to the adsorption of water and subsequentincrease in resin size (10,11). However, Am69 is a gel-typeresin with high volume and weight swelling. In this work,Dow88 had a weight swelling comparable to Am69 butsignificantly lower volume swelling. Both Dow88 and Am69possess sodium sulfonate ion exchange sites, which aredissociable and have a high affinity for water. However,Dow88 is a macroporous resin that has a higher degree ofcrosslink than gel-type resins (20). Thus, Dow88 had a similarpropensity for water adsorption to Am69; however, Dow88swelled slightly. This result suggested that matrix disintegra-tion created by embedded Dow88 might be the result ofwicking (water adsorption) rather than swelling actions.

Table I. Kinetic Parameters Analyzed from Release Profiles of HPMC-Based Matrices Containing Various Amounts of Dow88 or Am64

% Resin

Dow88 Am64

n k (h−n) R2 n k (h−n) R2

0 0.569 0.421 0.999 0.569 0.421 0.9995 0.547 0.386 0.999 0.504 0.412 0.99910 0.465 0.355 0.994 0.539 0.392 1.00020 0.422 0.260 0.977 0.492 0.372 0.99130 0.349 0.251 0.980 0.434 0.383 0.99340 0.227 0.239 0.977 0.410 0.395 0.991

Fig. 7. DPH release from EC-based matrices containing variousamounts of a Dow88 and b Am64, respectively

1111Comparison Between Strongly and Weakly Cationic Exchange Resins

Figure 7 displays the drug release profiles of EC-basedmatrices containing Am64. Due to the hydrophobicity of EC,Am64 addition caused non-swelling EC-based matrices todisintegrate; however, the extent of matrix disintegration wassignificantly less than observed in Dow88. Matrices contain-ing up to 20% Am64 disintegrated only slightly over 10 h, andthe resultant drug release profiles did not differ significantlyfrom those without the resin (Fig. 7). This result implied thatthe increase in drug release caused by disintegration wasoffset by the decrease in drug release due to drug exchangeon the embedded resin. As the amount of the embedded resinincreased from 30% to 40%, the extent of disintegrationincreased but was incomplete at 10 h, as confirmed byphotographic observation (Fig. 6c). Nonetheless, a greaterdisintegration of these matrices allowed for an increase indrug release, which dominated the adsorption of the drug onion exchange sites of the embedded resin. As a result, drugrelease was faster in matrices containing 30% to 40% Am64than from resin-free matrices. However, the amount of drugreleased from the matrix was likely less than resin-freematrices due to drug exchange and resinate formation. Dueto slow and incomplete disintegration, a portion of the drugwas bound to resin dispersed in the medium or resided in theremaining matrices. The decrease in drug release caused byAm64 was not as pronounced as the decrease caused byDow88 because of the weaker drug exchange property ofAm64. Thus, the profiles of drug release from EC-basedmatrices containing 30% to 40% Am64 were similar to thosecontaining up to 10% Dow88.

In general, it has been suggested in the literature thatsofter tablets disintegrate quickly, whereas harder tabletsdisintegrate more slowly. Therefore, pore structure is signifi-cant for tablet disintegration; softer tablets have larger poresand greater tortuosity, allowing for more water penetrationand subsequent breakage of inter-particle binding anddisintegration (4,5,25–27). Additionally, softer tablets haveweaker inter-particle bonds that are more easily broken bydisintegrating forces (3). EC-based matrices that containedincreasing amounts of Dow88 had reduced hardness (Fig. 2b)and increased pore size, as observed in SEM results (Fig. 3dversus e). Therefore, the decrease in hardness of EC-basedmatrices containing Dow88 might contribute to increaseddisintegration. On the other hand, Am64 yielded harder EC-based matrices due to the high compressibility of the resin(Fig. 2b), which needed to be overcome by disintegratingforces derived from the resin. Moreover, EC-based matricescontaining Am64 likely had smaller pores and greatertortuosity due to a denser powder bed (Fig. 3e versus f) thatimpeded water penetration into matrices. These factorsoutweighed the disintegrating forces caused by swelling andwicking (water adsorption) actions by Am64, leading todelayed matrix disintegration. Also, it explained that EC-based matrices containing Am64 did not disintegrate as muchas those containing Dow88 even though Am64 displayedhigher volume and weight swelling.

Drug Release from Matrices Under SimulatedGastrointestinal Conditions

According to the above results, resins had the potentialto slow down drug release from HPMC-based matrices. EC-

based matrices containing either resin had a propensity fordisintegration; thus, these matrices could not be used to delaydrug release. Therefore, the drug release of selected HPMC-based matrices containing Dow88 or Am64, as well as resin-free matrices, was further investigated under simulatedgastrointestinal conditions. Drug release from resin-freematrices under simulated gastric conditions was lower than

Fig. 8. DPH release from HPMC-based matrices containing a noresin, b 20%, and c 40% each resin tested in deionized water (opensymbols) and under simulated gastrointestinal conditions (filledsymbols), respectively

1112 Akkaramongkolporn, Wongsermsin, Opa nasopit and Ngawhirunpat

in deionized water (Fig. 8a). This result might be due to asalting-out effect caused by inorganic salts, which decreasedpolymer swelling and erosion, thus delaying drug release(7,8). This explanation was supported by photographs (Fig. 6)in which the gelled matrices under simulated gastrointestinalconditions were more persistent than in deionized waterduring drug release. Results revealed that Dow88 did notretard drug release under simulated gastrointestinal condi-tions as much as in deionized water, especially when theamount of embedded resin was high (Fig. 8b, c). This resultmight be due to competition between counter ions and thedrug for binding sites on the embedded resin (10,11). Thus,more drug was available due to a decrease in free binding sitesand was released from the matrices into solution. Matricescontaining a higher amount of embedded resin provided ahigher drug release because these matrices contained a loweramount of matrix former. Thus, the free drug was releasedfaster from the resultant matrix gel (5). In addition, the drugreleased in the buffer stage (2–10 h) was markedly higher thanthe amount released in the acidic stage (0–2 h). This wasbecause Na+, apart from H+, was present in the buffer stage,which had a higher affinity for strongly cationic exchangeresins, such as Dow88, compared to H+ (20). Additionally, thetotal amount of cations in the buffer stage (150 mEq of Na+

plus 75 mEq of H+) was three times higher than in the acidicstage (75 mEq of H+). Cations in the buffer stage couldcompete with the drug to a greater extent for binding sites onthe resin (28). In contrast, drug release from matricescontaining Am64 in deionized water was higher than undersimulated gastrointestinal conditions, in which both H+ andNa+ were present (Fig. 8b, c). This result suggested that bothH+ and Na+ could slightly compete with the drug in binding tothe resin under simulated gastrointestinal conditions.Therefore, the increase in drug release caused by competitionwith cations was minor and was dominated by a decrease indrug release due to the salting-out effect of the polymer, thusproviding a lower drug release under simulated gastrointestinalconditions. As shown in Fig. 8a–c, drug release from thematrices continuously decreased as the amount of Am64increased. Thus, Am64 could retard drug release undersimulated gastrointestinal conditions.

CONCLUSION

Dow88 and Am64 yielded matrices with unique physicalproperties. Dow88 caused the matrices to become softer,resulting in an increase in thickness, diameter, and friability.In contrast, Am64 increased matrix hardness and yielded athickness, diameter, and friability similar to resin-free matri-ces. The addition of Dow88 decreased drug release fromHPMC-based matrices in deionized water more than Am64.However, Am64 impeded drug release under simulatedgastrointestinal conditions. EC-based matrices containingeither resin had a propensity for disintegration caused byswelling and wicking (water adsorption) actions by the resin.

ACKNOWLEDGEMENTS

The authors wish to thank the Thailand Research Fundand the Commission on Higher Education, Ministry ofEducation, Thailand, for funding (MRG5280242). The

authors would also like to thank the Department of Pharma-ceutical Technology, Faculty of Pharmacy, Silpakorn Univer-sity, Nakhon Pathom, Thailand, for instrument support andthe Faculty of Medicine, Srinakharinwirot University, Thai-land, for assistance with SEM analysis.

REFERENCES

1. Kunin R. Ion exchange resins. New York: Wiley; 1963.2. Borodkin S. Ion exchange resins and sustained release. In:

Swarbick J, Boylan JC, editors. Encyclopedia of pharmaceuticaltechnology, vol. 8. New York: Marcel Dekker; 1993. p. 203–16.

3. Bandelin FJ. Compressed tablets by wet granulation. In: Lieber-man HA, Lachman L, Schwartz JB, editors. Pharmaceuticaldosage forms: tablets, vol. 1. New York: Marcel Dekker; 1989. p.131–94.

4. Sanchez-Lafuente C, Faucci MT, Fernandez-Arevalo M,Alvarez-Fuentes J, Rabasco AM, Mura P. Development ofsustained release matrix tablets of didanosine containing meth-acrylic and ethylcellulose polymers. Int J Pharm. 2002;234:213–21.

5. Dabbagh MA, Ford JL, Rubinstein MH, Hogan JE. Effects ofpolymer particle size, compaction pressure and hydrophilicpolymers on drug release from matrices containing ethylcellu-lose. Int J Pharm. 1996;140:85–95.

6. Crowley MM, Schroeder B, Fredersdorf A, Obara S, Talarico M,Kucera S, et al. Physicochemical properties and mechanism ofdrug release from ethyl cellulose matrix tablets prepared bydirect compression and hot-melt extrusion. Int J Pharm.2004;269:509–22.

7. Abrahamsson B, Alpsten M, Bake B, Larsson A, Sjogren J. Invitro and in vivo erosion of two different hydrophilic gel matrixtablets. Eur J Pharm Biopharm. 1998;46:69–75.

8. Kavanagh N, Corrigan OI. Swelling and erosion properties ofhydroxypropyl methylcellulose (hypromellose) matrices-influ-ence of agitation rate and dissolution medium composition. IntJ Pharm. 2004;279:141–52.

9. Kim JE, Kim SR, Lee SH, Lee CH, Kim DD. The effect of poreformers on the controlled release of cefadroxil from a polyur-ethane matrix. Int J Pharm. 2000;201:26–36.

10. Sriwongjanya M, Bodmeier R. Effect of ion exchange resins onthe drug release from matrix tablets. Eur J Pharm Biopharm.1998;46:321–7.

11. Akkaramongkolporn P, Ngawhirunpat T, Nunthanid J, Opana-sopit P. Effect of a pharmaceutical cationic exchange resin on theproperties of controlled release diphenhydramine hydrochloridematrices using Methocel K4M or Ethocel 7cP as matrix formers.AAPS PharmSciTech. 2008;9:899–908.

12. Jack DB. Handbook of clinical pharmacokinetic data. GreatBritain: Macmillan Publisher; 1992.

13. Reynolds JEF. Martindale the extra pharmacopoeia. 31st ed.London: The Royal Pharmaceutical Society of Great Britain;1996.

14. Ozyazici M, Gokce EH, Ertan G. Release and diffusionalmodeling of metronidrazole lipid matrices. Eur J Pharm Bio-pharm. 2006;63:331–9.

15. The United States Pharmacopeial Convention. USP 29. Rock-ville: The United States Pharmacopeial Convention; 2006.

16. Mamo M, Ginting D, Renken R, Eghball B. Stability of ionexchange resin under freeze-thaw or dry-wet environment. SoilSci Soc Am J. 2004;68:677–81.

17. Bajpai SK, Sharma S. Investigation of swelling/degradationbehaviour of alginate beads crosslinked with Ca2+ and Ba2+

ions. React Funct Polym. 2004;59:129–40.18. Dowex product literature. http://www.dow.com. Accessed 15 Jan

2010.19. Amberlite product literature. http://www.rohmhaas.com. Accessed

15 Jan 2010.20. Harland CE. Ion exchange: theory and practice. UK: Royal

Society of Chemistry; 1994.

1113Comparison Between Strongly and Weakly Cationic Exchange Resins

21. Kim CJ. Controlled release dosage form design. USA: Techno-mic Publishing Company; 2000.

22. Levina M, Rajabi-Siahboomi AR. The influence of excipients ondrug release from hydroxypropyl methylcellulose matrices. JPharm Sci. 2004;93:2746–54.

23. Avachat A, Kotwal V. Design and evaluation of matrix-basedcontrolled release tablets of diclofenac sodium and chondroitinsulphate. AAPS PharmSciTech. 2007;8:E1–6.

24. Zhao N, Augsburger LL. Functionality comparison of 3 classesof superdisintegrants in promoting aspirin tablet disintegrationand dissolution. AAPS PharmSciTech. 2005;6:E634–40.

25. Rudnic EM, Kottke MK. Tablet dosage forms. In: Banker GS,Rhodes CT, editors. Modern pharmaceutics third edition,

revised and expanded. New York: Marcel Dekker; 1996. p.333–94.

26. Lahdenpaa E, Niskanen M, Yliruusi J. Crushing strength,disintegration time and weight variation of tablets compressedfrom three Avicel PH grades and their mixtures. Eur J PharmBiopharm. 1997;43:315–22.

27. Riippi M, Antikainen O, Niskanen T, Yliruusi J. The effect ofcompression force on surface structure, crushing strength,friability and disintegration time of erythromycin acistratetablets. Eur J Pharm Biopharm. 1998;46:339–45.

28. Sprockel OL, Prapaitrakul W. Effect of eluant properties on drugrelease from cellulose acetate butyrate-coated drug resin com-plexes. Int J Pharm. 1988;48:217–22.

1114 Akkaramongkolporn, Wongsermsin, Opa nasopit and Ngawhirunpat